There are over 1 million persons in the United States who are intravenous drug users (IDUs). The epidemic has principally struck the minority inner-city population, with women affected nearly as frequently as men. Provision of medical services to IDUs places a significant strain on public resources, as close to 50 percent of this population lack any type of health insurance. Development of efficient diagnostic and management strategies for the common medical complications which occur in this group of patients have lagged behind that for other sectors of the population. Infective endocarditis (E) is one of the most serious medical complications associated with IDU, with significant morbidity and mortality occurring in unrecognized cases. Presenting clinical features associated with IE have been difficult to define. Because fever is found to occur in nearly 98 percent of cases of IE in IDUs, current practice is to hospitalize all febrile IDUs for further evaluation, testing and observation. Although the true incidence (approximately 3,000-7,500) of IE in IDUs is relatively low, nearly 10 times as many patients in whom the diagnosis is being considered are hospitalized each year. This represents a significant financial burden to the health care system. To date, there have been no studies which have attempted to develop alternative approaches to inpatient diagnosis in patients in whom the disease is being considered. The objectives of this study are: (1) To design an algorithm which uses clinical information and diagnostic adjuncts available during ED evaluation to identify which febrile IDUs are most likely to have IE, or are likely to develop complications that require hospitalization. IDUs with fever will be admitted from the ED to an in-patient unit arranged to simulate an ED observation unit. (2) To determine the potential costs savings associated with alternative management strategies (such as outpatient testing and follow-up, or ED based rapid evaluation) for those IDUs who have low risk for either of the above. Results from this pilot investigation will be used to design a larger prospective investigation to test the new clinical decision algorithm. The study will provide the first description of early use of diagnostic testing and clinical assessment applied in conjunction towards rapid evaluation of the large population of febrile IDUs. Implementation of this algorithm has the potential to save the health care system hundreds of millions of dollars each year in unnecessary hospital costs.